“The switching power supply has very high conversion efficiency and has become the mainstream product of the power supply. At the same time, loop analysis test is used more and more as its important test method. This article will introduce the principle and application of this test method in depth.
The switching power supply has very high conversion efficiency and has become the mainstream product of the power supply. At the same time, loop analysis test is used more and more as its important test method. This article will introduce the principle and application of this test method in depth.
Switching power supply is a typical feedback control system, which has two important indicators: response speed and stability. Response speed is the speed at which the power supply can quickly adjust when the load changes or the input voltage changes. Because the load of the switching power supply is mostly a digital IC, its current will change with the change of the logic function. For example, when the FPGA is configured, the current will more than double. The input voltage of the switching power supply will also fluctuate to a certain extent. In order to ensure the stable output of the power supply without drop or overshoot, it is required that the power supply must be adjusted quickly so that the final output voltage does not change. The response speed of the power supply determines the adjustment speed of the power supply.
Oscillation may occur due to the addition of a feedback system to the power supply. If the parameters of the power system are not set properly, oscillation will occur, and as a result, a fixed frequency fluctuation will be superimposed on the voltage. lead to unstable power supply.
The switching power supply is shown in the figure below:
Figure 1 Switching power supply debugging
It can be seen from the block diagram of the switching power supply that the system feeds back the final output change to the proportional circuit through a feedback circuit, and is input into the error amplifier through the proportional attenuation of the proportional circuit. Then, by comparing the difference between the signal and the internal reference signal, the error amplifier drives a series of output links such as the post-stage pulse width modulator, and finally cancels each other with the interference signal, thereby ensuring the stability of the power supply.
How should we measure the response speed and stability of the power supply? In the early debugging, we will use a variable Electronic load for testing, but because the frequency of change of the current electronic load is much lower than that of the switching power supply frequency, this method is gradually not used by everyone. At present, the more common test method is the loop test method. The loop test method is to inject a single frequency sine wave sequence signal into the feedback loop, and then judge its ability to adjust the interference of each frequency according to the output of the power system. The higher the gain of its loop response, the stronger the anti-interference ability of the power supply to this frequency band.
Figure 2 Loop test block diagram
The loop test block diagram is shown below:
As can be seen from the above figure, the loop test actually injects the interference signal into the error amplifier through the feedback circuit, and then checks the cascade response of the error amplifier and the output link of the latter stage. The response of the error amplifier is actually the open-loop gain of the error amplifier. So the fundamental purpose of the loop is shown in the following figure:
Figure 3 The fundamental purpose of loop testing
With the scanning of frequency signals one by one, the loop gain of each channel is finally drawn on a graph, and a very intuitive frequency domain characteristic graph will be obtained.
The final loop characteristic curve is shown in the following figure:
Figure 4 The corresponding curve of the loop
According to this picture, we can judge whether the power supply design is stable and whether there is room for optimization. The criteria for determining the stability of the curve are as follows:
Crossover frequency: It is recommended to be 5% to 20% of the switching frequency. If it is too high, it will be unstable, and if it is too low, the response speed will be too slow.
Phase margin: The requirement must be greater than 45°, and 45° to 80° is recommended.
Crossover slope (near 0dB): Single-pole crossover is required. Generally, the crossover slope is required to be around -1, that is, -20db/per decade.
Gain margin: It is recommended to be greater than 10dB.
The principle of interference signal injection
How to inject the interference signal into the error amplifier? The open-loop gain of the error amplifier is very large, about 60db. Then in order not to saturate the error amplifier output, the input signal must be around -50dbm, about 2mv. The amplitude of this signal is too small and it is too difficult to generate. The general electromagnetic noise signal must be higher than the amplitude of this signal. Obviously this direct injection is not feasible. In order to be able to successfully inject interfering signals, we need to use feedback to do so.
In order to explain the input method of the interference signal, we must first explain the feedback.
What is feedback? In addition to mathematical formulas, this concept is difficult to describe. You can think about an inverted pendulum that you often played with when you were a child. The eyes, brain, and hands form a feedback system. The eyes will feed back to the brain according to the swing of the stick, and the brain will control the hand to move, so as to keep the stick upright. As shown below:
This system is a very simple feedback system. The eyes correspond to the feedback circuit and the proportional circuit, the brain corresponds to the amplifier, and the nerves and hands correspond to the output stage such as the pulse width modulator.
Imagine that we add a filter between the eyes and the wooden stick in the above system. The only function of this filter is to make the image of the wooden stick sway left and right at a fixed frequency. After the eyes receive this swinging scene, they send signals to the brain, and in order to maintain balance, the brain starts to operate the hands to swing left and right. In the end, it looks like the wooden stick is already swinging in a small range in the scene after the filter, but in fact the wooden stick is swinging left and right, and its swing just cancels the effect of the filter. The ratio of the actual swing of the stick to the small swing in the scene is the adjustment capability of the entire system.
If you understand the hypothetical experiment above, it should be understood that we can string the signal between the feedback circuit and the proportional circuit by disconnecting the feedback circuit from the proportional circuit, just like that filter. The injection source converts the interference signal into a current signal through an isolation transformer, and generates an additional voltage difference Vfg across the injection resistor. Due to the virtual short characteristic of the negative feedback of the op amp, the error amplifier will adjust the output as much as possible at this time. Make the voltage of the positive and negative terminals of the op amp equal. This will eventually generate a ΔVout at the output stage, which is used to offset the extra voltage difference Vfg across the injection resistor. If the open-loop gain of the error amplifier is infinite, Vfg will be exactly equal to ΔVout. But since the open-loop gain of the error amplifier is limited, it will eventually lead to a ΔVin, ΔVin = Vfg – ΔVout. The ratio of ΔVin to ΔVout is the gain of the entire loop.
How to choose the injection point
There is a relatively simple way to select the injection point. For the voltage source, it is to find the two resistors used to calculate the voltage when designing the circuit. When designing the circuit, which resistor is used to adjust the output, it is added to which resistor. For the current source, it is roughly the same as the voltage source, but there is generally no R1 or R2 in the current source, as long as the injection resistor is placed after the feedback circuit.
The power loop test can clearly and accurately test the stability and response speed of the switching power supply, which is of great significance to the power supply design. The test method is also relatively simple. However, in the actual test, since the switching power supply system will generate a large number of harmonics, these harmonics will seriously affect the final test result, so the general oscilloscope cannot get more accurate results. The ZDS4000 oscilloscope adopts advanced FFT and FIR functions to reduce the interference of harmonics to a small extent, and at the same time greatly improves the resolution and accuracy of the loop curve. It’s no less effective than a dedicated power loop tester.